Why financial institutions need role-specific AI capability, not one-size-fits-all training

On 2 August 2026, the EU AI Act enters another major phase of application.

The EU AI Act is the European Union’s risk-based framework for governing how artificial intelligence is developed, provided and used. It applies to public and private organisations inside and outside the EU when they place AI systems on the EU market, introduce them into service or use them within the EU. This means the Act is relevant not only to technology companies developing AI, but also to organisations using it.

For financial institutions, this may include AI used in customer service, compliance, financial crime, risk management, credit assessment, investment analysis and employee productivity.

From August 2026, transparency requirements for certain AI systems and AI-generated content become applicable. National authorities will also begin supervising and enforcing the AI literacy obligation that has applied since February 2025. Although some deadlines for high-risk AI systems have recently been delayed, the August transparency and AI literacy enforcement milestones remain highly relevant to organisations using AI.

The question for financial institutions is therefore no longer simply whether they have an AI policy.

It is whether their people have the knowledge and professional judgement required to put that policy into practice.

AI literacy is already an organisational responsibility

Article 4 of the EU AI Act has applied since 2 February 2025.

It requires providers and deployers of AI systems to take measures to ensure that employees and others operating AI on their behalf have a sufficient level of AI literacy.

The appropriate level should reflect their technical knowledge, experience, education and training, as well as the context in which an AI system is used and the people it may affect.

This has an important practical implication:

AI literacy cannot be treated as a single introductory course delivered identically across an organisation.

A senior leader deciding where AI should be deployed needs different capabilities from a compliance professional assessing an AI-enabled process.

A product manager integrating AI into a customer journey has different responsibilities from an employee using generative AI to draft, research or analyse information.

They all require AI literacy, but they do not require the same AI literacy.

Why general AI awareness is not enough

General awareness training can create a useful shared foundation, but it may not equip people to carry out their specific responsibilities.

Knowing what generative AI is does not necessarily mean that an employee can recognise an unreliable output, protect confidential information or determine when human review is required.

Understanding basic AI concepts does not automatically enable a manager to assess whether an AI use case is appropriate for a regulated workflow.

Completing a general course may not equip a compliance professional to challenge a technology provider, identify governance gaps or assess whether human oversight is meaningful.

The European Commission states that there is no one-size-fits-all approach to AI literacy. Organisations should consider the systems they provide or deploy, the risks associated with them and the existing knowledge and responsibilities of the people interacting with them.

Without role-specific capability, organisations may experience:

• • Inconsistent or unreliable use of AI affecting business decisions

• • Weak human oversight, fragmented accountability and difficulty challenging technology providers

• • Exposure of sensitive information and increased regulatory and reputational risk

Regulation alone cannot create operational readiness. Organisations need people who can use, evaluate, challenge and oversee AI within the context of their responsibilities.

Build role-specific AI literacy with CFTE

Financial institutions need a shared foundation of AI knowledge, supported by deeper learning for professionals applying and governing AI.

CFTE provides three complementary pathways designed around different roles and responsibilities.

No individual course can guarantee compliance with the EU AI Act. Compliance depends on an organisation’s systems, risks, controls and governance. However, role-relevant learning can help build the knowledge and professional judgement needed to translate AI policies into practice.

Build essential foundations with AI Literacy for Professionals

AI Literacy for Professionals helps professionals understand, evaluate and confidently use AI while recognising its capabilities and limitations.

The programme develops foundational knowledge, critical evaluation of AI outputs, awareness of risks and greater confidence engaging with AI tools. It is suited to professionals building essential AI capabilities and organisations introducing workforce-wide AI literacy.

The outcome is not simply greater awareness. It is a workforce better equipped to make informed and responsible decisions about when and how AI should be used.

Apply generative AI with Generative AI 360 for Financial Services

Generative AI 360 for Financial Services helps professionals understand how generative AI is changing finance and how it can create practical value.

The programme combines expert-led learning with financial-services applications, helping business, product, operations, innovation and transformation professionals identify credible use cases and contribute more effectively to AI adoption.

Learners move beyond general awareness towards an applied understanding of how generative AI can support financial-services activities.

Strengthen governance with Generative AI for Compliance

Generative AI for Compliance in Financial Services is designed specifically for compliance professionals and other teams responsible for risk, governance and oversight.

Using compliance-focused applications and case studies, the programme explores regulatory reporting, risk identification, regulatory change and ethical AI practices. It helps professionals evaluate AI-enabled processes and engage more confidently with business and technology teams.

The outcome is stronger professional challenge and more informed oversight when AI is introduced into regulated processes.

One requirement, three complementary pathways

The EU AI Act creates an organisation-wide need for AI literacy, but the appropriate pathway depends on each person’s responsibilities.

AI Literacy for Professionals establishes the foundations required to understand, evaluate and use AI responsibly.

Generative AI 360 for Financial Services develops applied understanding of generative AI across finance.

Generative AI for Compliance in Financial Services builds specialist capabilities for compliance, risk and governance professionals.

Together, the programmes help institutions establish shared foundations, build practical capabilities and strengthen informed oversight.

Why CFTE?

CFTE develops applied AI capabilities specifically for financial-services professionals and institutions.

Its programmes combine AI knowledge with financial-services applications, professional responsibilities and practical implementation. CFTE’s AI in Finance Academy has supported the training of more than 20,000 professionals globally.

Prepare your workforce for responsible AI adoption

Preparing for the EU AI Act requires more than adding a general AI module to mandatory training.

Financial institutions need to understand who interacts with AI, which systems they use, what risks those systems create and what level of knowledge each group requires.

CFTE works with institutions to align AI learning pathways with workforce roles, organisational priorities and the context in which AI is being used.

Individual professionals can also select the pathway that best reflects their responsibilities:

• • Start with AI Literacy for Professionals to build essential knowledge and responsible-use capabilities.

• • Choose Generative AI 360 for Financial Services to understand and apply generative AI across finance.

• • Choose Generative AI for Compliance in Financial Services to strengthen compliance, risk, governance and oversight capabilities.

The question is no longer whether financial institutions will use AI.

It is whether their people are equipped to use, challenge and oversee it effectively.

Explore CFTE’s AI learning pathways or speak with CFTE about building role-specific AI literacy across your organisation

The post The EU AI Act’s August 2026 Milestone: Is Your Workforce AI Literate? appeared first on CFTE.

KEY TAKEAWAYS

• Global Women in AI officially launched its Hong Kong chapter on 9 July 2026 at LEAP East, Hong Kong Convention and Exhibition Centre.
• Only a small fraction of AI leadership positions globally are held by women. GWAI exists to change that permanently.
• The launch featured a recorded address from Chairwoman Tram Anh Nguyen and a panel of four distinguished women leaders from finance, fintech, and AI.
• GWAI is building toward 10,000 supercharged women AI leaders across 100+ countries through education, mentorship and community.

The Real Question Is Not Who Will Use AI.

It Is Who Will Shape It.

Artificial intelligence is transforming every industry.

From finance and healthcare to education, government and entrepreneurship, AI is changing how decisions are made, how businesses operate and how societies evolve.

But one important question remains:

Who will shape the future of AI?

Today, women remain significantly underrepresented in AI leadership and decision-making roles.

This is not only a diversity challenge. It is a leadership challenge.

If the people building and governing AI do not represent the world these systems serve, we risk creating technology that reflects only a limited set of perspectives.

As Tram Anh Nguyen, Chairwoman of Global Women in AI and Co-Founder of CFTE, highlights:

“The real question is not only who will use AI. The real question is who will shape the future of AI.”

Tram Anh Nguyen: A Call for Women to Lead AI

At the launch of the Hong Kong Chapter, Tram Anh Nguyen will share her vision for a future where women are not only participating in AI, but actively shaping its direction.

Through GWAI, the goal is clear:

To equip women with the knowledge, confidence and networks needed to become leaders, decision-makers and innovators in AI.

Why Hong Kong. Why Now.

Hong Kong is a strategic location for GWAI’s expansion.

As a global financial centre and a growing hub for technology and innovation, Hong Kong sits at the intersection of finance, entrepreneurship and AI transformation.

Launching the Hong Kong Chapter at LEAP East represents a commitment to ensuring that women across Asia have a voice in the global AI conversation.

The chapter will become a platform for:

• Learning and capability building
• Meaningful connections and mentorship
• Leadership opportunities across industries

Meet the Leader Behind GWAI Hong Kong

Jessica Liu
CEO and Co-Founder, Planto
GWAI Executive Committee & GWAI Ambassador and Lead, Hong Kong Chapter

The Hong Kong Chapter will be led by Jessica Liu, whose experience across entrepreneurship, fintech and innovation reflects the type of leadership GWAI seeks to cultivate.

Jessica’s role is to build a community where women across Hong Kong can connect, learn and contribute to shaping the future of AI.

Women Leaders Joining the Conversation

The launch event will feature a panel discussion bringing together inspiring women leaders from finance, fintech and AI.

The conversation will explore:

• How AI is reshaping industries
• What leadership looks like in the AI era
• How women can actively participate in building and governing AI systems

Panel speakers include:

Jessica Liu
CEO and Co-Founder, Planto
GWAI Executive Committee & GWAI Ambassador and Lead, Hong Kong Chapter
Moderator

Xiaoyan Cheyenne Shu
Ex Principal Investcorp Founder, AI Infrastructure Investment Company

Jessica Lam
Group Chief Strategy Officer, WeLab

Michele Fung
Head of APAC, Unlimit
Board Member, Fintech Association of Hong Kong

Building a Global Movement

Global Women In AI is more than a community.

It is a global movement focused on three pillars:

Education

Building strong foundations in AI, digital skills and systems thinking across every sector and career stage.

Mentorship

Creating access to role models, networks and communities that help women grow their confidence and leadership.

Leadership

Supporting women as decision-makers, founders, investors, policymakers and innovators shaping AI.

The ambition is to build a global network of 10,000 supercharged women AI leaders across 100+ countries.

Join the Hong Kong Chapter Launch

The future of AI will be shaped by those who step forward today.

Join Global Women in AI at LEAP East as we celebrate the launch of the Hong Kong Chapter and bring together leaders committed to building a more inclusive AI future.

9 July 2026
Hong Kong Convention and Exhibition Centre

https://luma.com/k9tr14sl

The post Global Women in AI Launches Hong Kong Chapter at LEAP East 2026 appeared first on CFTE.

Today, CFTE hosted a webinar with Dr David R. Hardoon, PhD, CFTE Senior Advisor, Data and AI leader, and former Chief Data Officer at the Monetary Authority of Singapore, on “Agentic AI in Financial Services: From Governance to Operationalisations.”

The session focused on one of the most important questions facing financial institutions today:

How do we move from responsible AI principles and governance frameworks to practical implementation, assurance and controls?

As generative AI and AI agents become more embedded in financial services workflows, the challenge is no longer only about understanding the technology. It is about building the right operating model for responsible adoption.

The problem is not a lack of governance

Financial services is not short of governance.

There are frameworks, policies, principles, model risk processes, regulatory expectations, internal controls and approval committees. Yet, when organisations try to deploy AI, especially agentic AI, governance is often experienced as a blocker rather than an enabler.

During the webinar, participants identified familiar barriers to agentic AI adoption: governance, security, lack of expertise, data quality, control and awareness.

Dr Hardoon’s point was not that governance is unnecessary. Quite the opposite.

The intended purpose of governance is to create safeguards, manage risk, build trust, improve transparency and support long-term value. The problem is that, in practice, governance can become red tape, box-ticking and a process that slows down deployment without necessarily improving outcomes.

That gap matters.

If governance already feels difficult in traditional AI, agentic AI increases the complexity. Instead of one model producing one output, organisations may now be dealing with multiple agents, tools, workflows, decisions, escalations and human interactions.

This is why governance cannot remain a one-time approval gate.

Agentic AI is about orchestration

A key theme from the session was that agentic AI should not be understood as “AI acting alone.”

Dr Hardoon described agentic AI as an environment of specialised agents, task decomposition, memory, orchestration and collaboration. Some agents may perceive, some may reason, some may act, and others may coordinate.

But in financial services, the most important point is this:

Agentic systems will not consist only of AI agents.

They will sit alongside human agents, rule-based systems, existing operational processes and business teams. A true agentic environment is therefore not only a technical architecture. It is an operating model.

This creates a governance challenge.

If institutions govern AI agents in one way, rule-based automation in another way and human decision-making in another way, they risk creating inconsistent controls across the same business process.

The governance question becomes less about whether something is “AI” and more about what role it plays in the business process, what risk it introduces, and how it should be monitored.

The double standard problem

One of the strongest examples from the webinar was the idea of double standards.

When an AI solution accesses files, analyses information and drafts a memo, cybersecurity, legal or compliance teams may raise concerns. Those concerns may be valid.

But the next question is: how is the same task done today?

If a human already accesses the same files, reviews the same information and drafts the same memo, then the issue is not simply that “AI creates risk.” The issue is whether the new process changes the existing risk profile.

Does it increase risk?
Does it reduce risk?
Does it create new controls?
Does it expose gaps in the way the existing process was already governed?

This is where financial institutions need a more mature approach. AI should not be treated as special simply because it is AI. It should be governed according to the role it plays, the risk it introduces and the outcome it is expected to produce.

From one-time approval to continuous governance

Traditional governance often assumes a deterministic process.

A system is reviewed, approved, deployed and then periodically checked. That works reasonably well when inputs and outputs are stable and predictable.

Agentic AI is different.

The system may adapt, interact with changing data, respond differently depending on context and operate across a wider workflow. In this environment, governance has to become continuous.

Dr Hardoon highlighted several dimensions institutions need to monitor.

Observability means being able to see and understand what the agent or fleet of agents is doing.

Controllability means having a way for a human, another agent or a rule-based system to intervene when needed.

Stability means ensuring the wider system can remain reliable if one agent fails or performs below expectation.

Robustness means testing whether the system can handle adversarial prompts, market shocks, noise, system issues or unexpected inputs.

Performance means checking whether the system is still delivering the intended business outcome.

This is a very different mindset from asking whether a model passed an approval gate once.

It is closer to running an operating environment.

Governance as a business function

One of the most practical takeaways from the session was the need to treat governance as a business function.

That means governance should not sit outside the development lifecycle as a separate approval step. It should be embedded into the process from the beginning.

A practical AI governance operating model might include intake, assessment, build and test, deployment approval, operate and monitor, and retire or decommission.

At the intake stage, the AI use case is submitted, reviewed and added to an inventory with clear ownership.

At the assessment stage, risk, compliance, privacy, cybersecurity and business impact are reviewed early.

During build and test, evidence, attestations and controls are collected during development, not after everything is built.

Deployment approval should then be based on the risk tier, business context and evidence gathered.

Once deployed, performance, drift, value, risk indicators and compliance need to be monitored continuously.

Finally, institutions also need to think about retirement and decommissioning, including end-of-life decisions, migration and replacement.

This is where governance becomes useful. It becomes part of execution, rather than a hurdle that appears at the end.

The human capital gap

The session also ended with an important reminder: AI is ultimately about people.

Dr Hardoon noted that HR is often missing from AI initiatives, despite the fact that AI directly affects skills, roles, capabilities, operating models and talent strategy.

This is particularly important for agentic AI.

If AI agents begin to work alongside human agents, organisations need to rethink more than technology. They need to rethink capability building, role design, accountability, workforce readiness and leadership.

AI governance cannot be separated from human governance.

The question is not only: “How do we govern AI?”

It is also: “How do we govern work in an environment where humans, AI systems and rule-based processes operate together?”

What financial institutions should do next

The lesson from the webinar is clear.

Financial institutions do not need more abstract AI governance. They need governance that can be operationalised.

That means moving from policies to processes.

It means moving from one-time approvals to continuous monitoring.

It means moving from AI-specific controls to unified governance across humans, rules and AI agents.

It means moving from experimentation to operating models.

And it means moving from technical adoption to capability building.

Agentic AI will make governance more important, not less.

But the institutions that succeed will be the ones that make governance practical enough to accelerate responsible adoption, rather than slow it down.

The post Agentic AI in Financial Services: Governance Has to Move From Policy to Operating Model appeared first on CFTE.

Reflections from CFTE’s conversation with Dr David R. Hardoon in Abu Dhabi

There is a familiar problem inside many organisations trying to use AI.

The business wants to move quickly.
The technology team wants to experiment.
The governance team wants to understand the risks.
Leadership wants confidence before anything is scaled.

Too often, these teams meet too late.

A use case is built. A pilot is tested. Then governance is asked to approve it. At that point, the conversation can feel like a stop sign: too risky, not clear enough, not ready, not aligned.

But governance should not be where innovation stops.

That was one of the key messages from CFTE’s recent conversation with Dr David R. Hardoon in Abu Dhabi: good governance should help organisations move forward, not hold them back.

The problem is not governance. It is when governance starts.

AI is now part of real work in financial services. It is being used in customer service, fraud detection, compliance, operations, software development, research, productivity, and decision support.

That means governance can no longer be treated as a final approval step.

If governance only comes in at the end, teams are left guessing what is acceptable. They may build something that cannot be used. They may move too slowly. Or worse, they may avoid the process altogether because it feels disconnected from the work.

The better approach is to bring governance in earlier.

Before a team starts building, they should already understand the basic questions:

What data can we use?
What risks are we comfortable with?
Where do we need human oversight?
What happens if the system makes a mistake?
Who is responsible?

These questions do not stop innovation. They make innovation usable.

From “no” to “how”

Dr Hardoon has been part of CFTE’s AI journey since 2018. Back then, the conversation around AI in finance was still mostly about machine learning, data, model governance, and responsible experimentation.

Today, the conversation has changed.

Generative AI has entered daily workflows. Agentic AI is raising new questions about autonomy and oversight. Regulators are looking closely at how institutions balance innovation with trust.

But one message has stayed consistent: governance should not be the department of “no”.

Its role should be to help teams understand how something can be done safely, where the boundaries are, and whether the organisation is ready to proceed.

That is a much more useful way to think about governance.

AI cannot be governed once and forgotten

One reason AI is difficult to govern is that it does not behave like a normal business process.

A traditional process can often be reviewed, approved, documented, and checked again later.

AI is different.

Models change.
Outputs vary.
Users behave differently.
New risks can appear after deployment.

This is especially true when AI systems produce different answers depending on prompts, context, data, or user behaviour.

So governance cannot be a one-time gate. It has to continue after launch.

That does not mean making every AI project slow or heavy. It means having practical ways to monitor what is happening, spot issues early, and know when someone needs to step in.

Done well, governance does not slow teams down. It prevents them from reaching the end of a project and discovering that what they built cannot actually be used.

Agentic AI makes this even more important

The conversation became especially interesting when it turned to agentic AI.

Agentic AI is often described as AI that can act more independently. But for organisations, the bigger issue is not just the AI agent itself. It is the system around it.

In a bank, an AI tool does not operate in isolation. It connects with policies, customer data, internal workflows, human teams, escalation rules, and compliance requirements.

Take a call centre as an example.

The question is not only: did the AI give the right answer?

The better questions are:

Was the answer based on the right information?
Did it follow the bank’s policy?
Did it know when to escalate to a human?
Can the organisation see what happened?
Who is accountable if something goes wrong?

This is why AI governance cannot only focus on the model. It has to look at the whole system.

Governance is also about people

Another important point from the discussion was that governance is not only about policies and controls.

It is also about capability.

Many organisations already have access to AI tools. That is not the main problem.

The bigger challenge is that different teams understand AI in very different ways.

Business teams want speed.
Technology teams want to test.
Risk teams need clarity.
Leaders need confidence.
Employees need to know what responsible use looks like.

When these groups do not share a common language, AI adoption becomes slow and fragmented.

This is why capability building is part of governance.

A policy only works if people know how to apply it. A framework only matters if teams can use it in real decisions. Responsible AI principles only matter if they change how people design, use, monitor, and challenge AI systems.

Why this matters in Abu Dhabi and the UAE

David’s visit to Abu Dhabi reflects a wider priority for CFTE: bringing global experts into the UAE and the Middle East to support the region’s ambitions in AI, finance, and future skills.

The region is not watching the AI conversation from the sidelines.

Governments, regulators, financial institutions, and enterprises are actively investing in AI as a driver of productivity, competitiveness, workforce transformation, and innovation.

But what makes the conversation especially important is the focus on people.

AI adoption is not only about infrastructure, models, or tools. It is about whether leaders, teams, and institutions have the skills and judgement to use AI well.

That is where global expertise needs to meet local ambition.

The right question is not simply: what is global best practice?

A better question is: what is globally informed, locally relevant, and useful in practice?

The human skill that matters most

One of the strongest ideas from the conversation was about human judgement.

As AI becomes more capable, people do not become less important. Their role changes.

People need to become better at asking questions, framing problems, recognising trade-offs, interpreting outputs, and taking responsibility for decisions.

David described cognitive skills as a muscle. That is a useful way to think about AI readiness.

If AI helps more with execution, then human value moves towards judgement.

What should we ask?
What should we trust?
What should we challenge?
What could go wrong?
What does a responsible decision look like?

These are the questions that will determine whether AI creates real value or introduces new risks.

The real work ahead

We are grateful to Dr David R. Hardoon for joining us in Abu Dhabi and for continuing to contribute to CFTE’s work.

His message is timely for financial services and for the region.

Governance and innovation should not sit on opposite sides of the table. Governance should help teams move with more clarity. Innovation should bring governance into the process earlier. Capability building should give people the confidence to make better decisions.

At CFTE, this is why we bring experts like David to the UAE and the wider Middle East.

The aim is not to add another conversation about AI trends. It is to support the people and institutions building the future of finance in practice.

AI governance is not the brake.

Done properly, it is how organisations learn to move faster, safer, and with greater confidence.

The post AI Governance Is Not the Brake. It Is the Way Forward. appeared first on CFTE.

The post UK Finance and CFTE Launch New Offering to Build AI Capability Across UK Financial Services appeared first on CFTE.

The UK’s AI objectives are changing. It is no longer only a question of ambition, innovation or adoption. It is now a question of implementation.

We are entering the age of agentic AI: systems that do not only assist, but act. They can plan, reason across steps, coordinate with other agents and operate inside the tools organisations already use. Generative AI changed how people create and analyse information. Agentic AI will change how work is organised, how decisions are made and how institutions deliver value.

For the UK, this creates a strategic opening. The country has world-class research, a strong technology ecosystem, sophisticated regulators, deep sector expertise and a government that has placed AI at the centre of national growth. But if we want AI to translate into productivity, trust and inclusive prosperity, the next phase cannot be only about tools, models or infrastructure. It must also be about people.

The UK does not only need AI adoption. It needs AI fluency.

Adoption means access to tools. Fluency means the capability to understand what AI can do, where its limits are, when to trust it, when to challenge it and how to apply it responsibly in context. It is the difference between using AI as a faster task-doer and knowing how to redesign work around intelligent systems.

This distinction matters because agentic AI is not just another software upgrade. When a system can act, leaders must ask new questions. What should be automated? Who authorises an action? Who is accountable for the outcome? How do we audit decisions made or supported by autonomous systems? When should humans intervene? What happens when an agent acts across multiple systems, suppliers or jurisdictions?

These are not purely technical questions. They are questions of operating model, governance, skills and trust.

The opportunity is already visible. In financial services, Lloyds Banking Group has announced an AI-powered financial assistant for millions of customers. In education, the UK government is supporting the development of AI tutoring tools that could help disadvantaged pupils. In public services, police forces are experimenting with digital agents to respond to non-emergency citizen queries and free human teams to focus on higher-value work.

The next frontier is agentic commerce, where agents search, compare, negotiate and buy on behalf of individuals and businesses. For companies, the question will shift from “How do we reach the customer?” to “How do we become discoverable, trusted and chosen by the agent acting for the customer?”

This is the scale of the change ahead. Agentic AI will begin to decouple cognitive work from headcount. Organisations will be able to increase output without increasing people in the same way. Some professionals will be outcompeted by systems that are faster, cheaper and always available. Others will be supercharged, using AI with their judgement, domain expertise and contextual understanding to reach a new level of performance.

The dividing line will not be the technology. It will be capability.

From our work with financial institutions, regulators and governments internationally, we see the same pattern repeatedly. Organisations rarely struggle because they lack access to AI tools. They struggle because their people do not yet have the fluency to redesign workflows, govern risks and create value responsibly.

This is why AI fluency must become a national priority.

If fluency remains concentrated among a small group of technical experts or already advantaged firms, AI will widen the gap between those who can benefit and those who are displaced. But if fluency is built deliberately across sectors, regions and communities, it can become a foundation for broader participation in the AI economy.

For the UK, this requires a human-capital infrastructure for AI.

First, the UK needs a shared understanding of AI fluency. Not everyone needs to become an engineer. But citizens, frontline workers, managers, executives, regulators and board members all need the ability to use, question and govern AI at the right level for their role. A common fluency framework would help organisations move beyond generic awareness and define the capabilities people actually need.

Second, fluency must be built around sectors and roles. AI in banking is not the same as AI in healthcare, education, legal services, manufacturing or public administration. The workflows, risks, accountability structures and professional norms are different. The future belongs to AI-bilingual professionals: bankers, clinicians, civil servants, teachers, regulators and entrepreneurs who understand both their field and how AI transforms it.

Third, capability must be embedded into responsible deployment. Organisations should not be assessed only on whether they use AI, but on whether they have the skills, governance, auditability and oversight to use it well. Procurement, regulation and institutional strategy should reward responsible implementation, not just technological novelty.

Fourth, inclusion must be designed into the system from the start. Women, underrepresented communities, workers outside major technology hubs, SMEs and professionals in roles most exposed to automation must not be treated as an afterthought. AI fluency cannot be a privilege for the few. It must become a capability distributed across the economy.

This is the human-capital layer of AI leadership that has been missing from too many AI strategies. At CFTE, our work has shown that responsible AI adoption is not a one-off training exercise. It is institutional readiness built as a system: defined through proficiency frameworks, delivered through role-based pathways tied to real workflows and measured through capability diagnostics.

Technology will continue to advance. Models will become more powerful, cheaper and more autonomous. But the countries that lead will not simply be those that move fastest. They will be those that build the deepest capacity to adapt, govern and create value responsibly.

For the UK, the real opportunity is not only to be a global AI innovator. It is to become a global model for responsible AI implementation.

That means recognising capability as infrastructure. It means treating AI fluency as seriously as compute, data and regulation. And it means ensuring that the people who power the economy are not passive recipients of technological change, but active shapers of it.

The future of agentic AI will not be determined by technology alone. It will be determined by people. The choice for the UK is whether we shape this shift, or are shaped by it.

The post Building the Human Infrastructure for the UK’s Agentic AI Economy appeared first on CFTE.

Today, Anthropic released Claude Fable 5 and Claude Mythos 5. This is not just another AI launch. It is one of the clearest signals yet that AI has entered a new phase, and financial services is directly in its path.

The most capable AI model ever made available to the public is now in everyone’s hands, more capable and far cheaper than the model before it. The single most important thing to understand is not the model itself. It is what the model signals. AI capability is now compounding faster than the governance frameworks, risk cycles, operating models, and skills that financial services was built on.

This article explains what launched, what a Mythos-class model is, why it matters for banks, regulators, and professionals, and what to do about it.

What did Anthropic actually launch?

Anthropic released two models. They are easy to confuse, and the difference matters.

Claude Fable 5 is the public model. It is a Mythos-class model that Anthropic has made safe for general use, and the company describes it as the most capable model it has ever made generally available. It is available to everyone today.

Claude Mythos 5 is not public. It is the same underlying model with its cybersecurity safeguards lifted, restricted to a small group of partners through Project Glasswing in collaboration with the US Government, with select biology researchers to follow.

So the headline matters: Mythos itself did not go public. A Mythos-class model, Fable 5, did. That distinction is the difference between a controlled frontier capability and a capability now sitting on every professional’s desk.

What is a Mythos-class model?

A Mythos-class model is a tier of capability that sits above Anthropic’s previous Opus class. The first, Claude Mythos Preview, was released in April 2026 to a limited set of cyber defenders and infrastructure providers because its capabilities were considered too powerful for open release. Fable 5 and Mythos 5 follow it.

According to Anthropic, Fable 5 is state of the art on nearly all tested benchmarks, with particular strength in software engineering, knowledge work, vision, and scientific research. The longer and more complex the task, the larger its lead over previous models. For finance specifically, Anthropic reports that Fable 5 tops Hebbia’s finance benchmark for senior level reasoning, with gains in document based reasoning, chart and table interpretation, and problem solving.

In plain terms: this is not an incremental upgrade. It is a step change in what a single professional can do.

Why is this the signal of exponential AI?

At CFTE we have been describing a shift we call the Third Wave of AI. It is the lens that makes today’s launch legible.

The first wave was machine learning, AI that could predict. The second wave was generative AI, AI that could reason, the moment most people met through ChatGPT. The third wave is AI that can act. It uses tools, runs multi step workflows, and executes in the digital world, not just answering questions but completing work.

Fable 5 is a clear marker of that third wave. Anthropic reports that an early tester, Stripe, used the model to perform a codebase wide migration across a fifty million line codebase in a single day, work that would otherwise have taken a team more than two months by hand.

Three forces are compounding this shift at the same time:

1. AI is becoming sharply better. Capability that was state of the art months ago is now routine, and the hardest, longest tasks are where the newest models pull furthest ahead.
2. AI is becoming sharply cheaper. Fable 5 launched at less than half the price of the model before it, at 10 dollars per million input tokens and 50 dollars per million output tokens.
3. AI is becoming sharply more autonomous. Models now sustain useful work across long, complex tasks rather than short bursts.

When capability rises and cost falls at the same time, year after year, the result is exponential, not linear. That is the real story. The model is only the signal.

What does exponential AI mean for financial services?

Financial services has invested in AI for years. The challenge now is not access to the technology. It is the capability to use it well, govern it, and turn activity into execution.

Three implications stand out.

Governance is now a moving target. Anthropic released Fable 5 to the public while deliberately gating Mythos 5 and routing sensitive cyber, biology, and chemistry queries to a more constrained model. A frontier lab choosing to release one model and restrict another is a live example of the governance question every regulated institution now faces. Risk frameworks built for annual cycles are being tested by capabilities that change month to month.

Performance is being redefined. When one professional with the right tools can do what once took a team, the definition of high performance changes. This is the heart of CFTE’s work on the AI-fication of Jobs and the emergence of what we call Supercharged Professionals, people who combine deep domain judgment with real AI leverage.

The workforce will split three ways at once. Using the CDE Innovation Prism, a framework for how technology reshapes work, we expect three patterns to happen simultaneously: displacement where AI competes directly with tasks, a small number of creative disruptors, and a growing class of supercharged professionals. The dividing line between them is not access to AI. It is the capability to use it well.

Prepare before the next Mythos moment

CFTE has been preparing financial services for this shift for years. We identified AI as the single most important technology impacting financial services in 2018, and have built the frameworks and programmes to turn that conviction into institutional capability ever since.

The gap that decides who thrives in exponential AI is capability, and capability can be built deliberately.

CFTE co-founder Huy Nguyen Trieu and Professor Douglas Arner of the University of Cambridge recorded a free 90 minute course, Exponential AI and the Impact on Financial Services, before today’s launch. It explains the Three Waves of AI, the forces compounding them, how regulators are responding, and how individuals and institutions can move from watching to building. It is free and includes a certificate.

Access the course and find more information here: https://courses.cfte.education/exponential-ai-and-the-impact-on-finance/

Frequently asked questions

Is Claude Mythos available to the public?

No. The public model is Claude Fable 5, a Mythos-class model made safe for general use. Claude Mythos 5, the version with cybersecurity safeguards lifted, is restricted to selected partners through Project Glasswing and the US Government.

What is exponential AI?

Exponential AI describes the pattern of AI becoming sharply more capable and sharply cheaper at the same time, year after year, so that what is impossible today can become routine in a matter of months rather than decades.

Why does the Fable 5 launch matter for finance?

Because the most capable AI model ever made public is now available to every institution and professional, at less than half the previous price, and it can complete complex work rather than only assist with it. This compresses the gap between large and small institutions and tests governance frameworks built for slower change.

Why does CFTE call this a signal of Exponential AI?

Because the release is not only about one model. It shows a broader pattern: AI capability is improving faster, becoming more widely available and affecting more complex tasks. That is the core idea behind Exponential AI.

The post Mythos Has Arrived: Why Financial Services Must Prepare for Exponential AI appeared first on CFTE.

Artificial intelligence is rapidly reshaping financial services. From credit scoring and insurance pricing to fraud detection, algorithmic trading, compliance and wealth management, AI is becoming embedded in the systems that influence how financial institutions operate and how customers access financial products.

But as AI becomes more powerful, one question is becoming increasingly urgent: how can financial institutions unlock the benefits of AI while ensuring that it remains safe, fair, explainable and accountable?

This was the focus of CFTE’s latest AI and Finance webinar, delivered in collaboration with Global Women in AI, featuring Dr Amira Abdelaziz, Senior Advisor for Data Science and Advanced Analytics at the Central Bank of Egypt and Executive Committee member of Global Women in AI.

In her session, “Beyond the Horizon: Singularity, Finance and the Governance Imperative,” Dr Amira explored how the next stage of AI could transform finance, why governance can no longer be treated as optional, and what institutions can do now to prepare.

Why AI Governance Matters in Financial Services

Dr Amira opened the session with a powerful scenario: by 2030, an AI system could manage savings, price insurance, approve loans and make financial decisions — while its reasoning remains invisible to the people affected by those decisions.

This is the core challenge of AI governance in finance.

Financial services is built on data, risk, trust and regulation. Every transaction, credit decision, compliance check, customer interaction and investment recommendation generates information. This makes finance one of the sectors most ready for AI adoption.

It also makes finance one of the sectors most exposed if AI systems go wrong.

As Dr Amira explained, AI governance is no longer a “nice to have”. It is becoming essential because financial institutions need to know how AI systems are being used, what decisions they influence, what data they rely on, how they are monitored, and who is accountable when something fails.

Finance Is AI’s Native Environment

One of the strongest messages from the webinar was that finance is a natural environment for AI because it is a pure information business.

Every decision in finance can be quantified. Every risk can be measured. Every transaction creates a signal. AI systems are designed to detect patterns, process data at speed and support decision-making at scale.

This is why AI is already transforming multiple areas of financial services.

In algorithmic trading, AI systems can process news, market signals, social media sentiment and price movements in real time. In insurance underwriting, AI can reduce decision-making from weeks to seconds by analysing risk variables, fraud indicators and behavioural patterns. In credit scoring, AI can use alternative data to assess customers who may not have traditional credit histories. In fraud detection, AI can identify suspicious transactions faster than rule-based systems. In wealth management, robo-advisors can make investment guidance more accessible to wider groups of customers.

These use cases show the opportunity. AI can improve speed, efficiency, access and personalisation across financial services.

But they also show why governance is critical.

The Risks Financial Institutions Cannot Ignore

AI in finance brings significant benefits, but it also introduces risks that institutions must manage carefully.

Dr Amira highlighted several governance risks that financial institutions need to address:

Algorithmic bias can occur when AI systems are trained on historical data that already contains discrimination. If past lending or hiring decisions were biased, an AI system may learn and amplify those patterns.

The explainability gap becomes a major issue when AI decisions cannot be clearly explained to customers, regulators, auditors or even internal teams. In financial services, this is especially important because decisions can affect access to credit, insurance, savings and investment opportunities.

Concentration risk can emerge when many financial institutions depend on a small number of AI vendors or models. A vulnerability, update failure or model weakness could become a systemic risk.

Adversarial AI creates new security challenges. Bad actors may attempt to manipulate AI systems, poison data or design fraud patterns that are difficult for automated systems to detect.

Job displacement and workforce disruption also need to be considered, particularly as AI systems become capable of performing increasingly complex tasks across compliance, operations, legal, customer service and analysis.

The lesson is not that financial institutions should avoid AI. The lesson is that they must adopt AI responsibly.

Explainability Is Becoming a Trust Issue

A recurring theme throughout the webinar was explainability.

If an AI system denies a mortgage application, prices insurance differently for two customers or flags a transaction as suspicious, the institution must be able to explain why. Customers need to be able to challenge decisions. Regulators need to audit systems. Internal teams need to understand how models behave.

Dr Amira used real-world examples to show what happens when explainability is weak. In cases where institutions could not explain why a model produced a certain outcome, the issue became more than a technical problem. It became a regulatory, reputational and trust problem.

For financial institutions, explainability is no longer only about model transparency. It is about customer protection, compliance and accountability.

AI Governance Needs to Be Built from the Beginning

One of the clearest messages from the session was that governance must be embedded before AI systems scale.

Institutions cannot wait until after deployment to ask whether a model is fair, explainable or secure. Governance needs to be part of design, procurement, implementation, monitoring and review.

Dr Amira outlined five practical pillars of AI governance:

1. Accountability
   Organisations need a named AI risk owner or AI governance leader with a real mandate, not just a general department responsible for AI.
2. Transparency
   AI decisions must be logged and explainable, especially when they affect customers.
3. Fairness
   Bias testing should not happen once. Models drift over time, so audits need to be repeated regularly.
4. Security
   AI models should be treated with the same security importance as core banking systems.
5. Human oversight
   Humans must be able to intervene in high-impact decisions, especially in areas such as credit, insurance pricing and anti-money laundering.

A Practical 90-Day Roadmap for AI Governance

For institutions wondering where to begin, Dr Amira shared a practical 90-day roadmap.

In the first 30 days, organisations should conduct a full AI inventory. This means identifying every AI system being used across the institution, including tools embedded in third-party vendor contracts.

In the next 30 days, they should classify AI systems by risk. High-risk areas such as credit scoring, anti-money laundering and insurance pricing require stronger oversight, documentation and explainability.

In the final 30 days, institutions should begin to act. This includes appointing an AI risk owner, drafting an AI policy, implementing explainability logging for high-risk decisions and scheduling regular model audits.

The key message was simple: you cannot govern what you cannot see.

Emerging Markets Need Their Own AI Governance Approaches

Another important insight from the webinar was the need for AI governance frameworks that reflect the realities of emerging markets.

Many global AI governance frameworks have been developed from high-income, Western regulatory contexts. They often assume strong digital identity infrastructure, mature credit bureau systems, advanced supervisory capacity and well-established legal mechanisms.

But many financial institutions and regulators in emerging markets operate in very different environments.

Dr Amira highlighted that this should not be seen only as a challenge. Emerging markets also have an opportunity to develop innovative governance models that reflect local needs, financial inclusion goals, multilingual environments, sparse data contexts and mobile-first financial ecosystems.

This is especially important as AI has the potential to expand access to finance for people who have historically been excluded from traditional banking.

What This Means for Finance Professionals

For CFTE, this webinar reinforced a key point: the future of finance will not only be shaped by technology, but by the people who understand how to use it responsibly.

AI governance is becoming a core capability for financial services professionals. It is relevant not only for data scientists and technology teams, but also for leaders, regulators, compliance professionals, risk teams, product managers, legal teams and customer-facing functions.

As AI becomes embedded across financial services, professionals will need to understand not only what AI can do, but how it should be governed.

Conclusion: The Future of Finance Requires Responsible AI

AI is already changing financial services. The next stage will bring even greater speed, scale and complexity.

The institutions that succeed will not simply be those with the most data or the most advanced models. They will be those with the strongest governance, the clearest accountability and the ability to use AI responsibly.

As Dr Amira reminded participants, the future of AI in finance is not something to fear. It is something to prepare for.

At CFTE, we are committed to helping financial institutions, regulators and professionals build the knowledge and capabilities needed for this next chapter of finance.

The post AI Governance in Finance: Why Responsible AI Is Now a Strategic Priority appeared first on CFTE.

Mythos alarmed regulators, central banks and bank leaders globally. The institutions treating it as a one-off event are misreading the situation. AI capability is compounding. Deployment cost is collapsing. Financial services was built for linear change. This post explains the shift, what it demands across four dimensions of readiness, and what professionals and institutions should do before the next wave arrives.

The Alarm Went Off. Most Institutions Pressed Snooze.

When Mythos was released, the reaction was immediate. The BBC, the New York Times, Reuters and the Financial Times all covered it. Regulators convened emergency meetings. Central banks issued warnings. The world’s most senior financial policymakers put it on the agenda.

And then, gradually, the conversation is losing its steam.

This is precisely the wrong response.

Mythos matters not because of what it is. It matters because of what it signals. A new class of AI systems that can reason, plan, use tools and act with increasing autonomy has arrived in financial services. If Mythos is the first visible signal, the question the industry should be asking is whether it is ready for the next ten.

At CFTE, we have been preparing financial services professionals and institutions for this moment since 2018, when we identified AI as the single most important technology reshaping the sector. What we are witnessing now is not a moment to observe. It is a moment to prepare.

Exponential AI: Why Linear Thinking Will Fail You

Most institutions are making a category error. They are tracking AI development the way they track everything else: reviewing quarterly model releases, updating internal policies in annual cycles, running pilot programmes before committing to adoption.

This is the wrong mental model.

AI capability is not growing linearly. It is compounding. Each generation of models does not add incremental capability. It multiplies it. Simultaneously, the cost of deployment has collapsed. What required enterprise infrastructure and seven-figure technology budgets two years ago is now accessible at a fraction of the cost.

The result is a structural asymmetry that financial services has not encountered before. The sector operates on regulatory cycles, planning horizons, risk frameworks and workforce development programmes calibrated for a predictable environment. Exponential AI is not predictable.

This is what CFTE calls the Exponential Preparedness Gap.

AI capability is advancing faster than institutions, regulators and professionals are adapting their governance, operating models, risk frameworks and skills. Without structural intervention, the gap widens with every capability jump.

The Three Waves of AI: Where We Are Now

The First Wave: Machine Learning and Prediction

The first wave gave institutions the ability to analyse large datasets and predict outcomes with greater accuracy. Credit scoring, fraud detection and risk modelling were transformed. AI as a tool that enhanced existing processes.

The Second Wave: Generative AI

The second wave gave professionals the ability to write, code, synthesise and assist at scale. This wave was characterised by AI as a collaborator in professional work.

The Third Wave: Agentic and Autonomous AI

This is the wave Mythos represents. Systems that can reason across multiple steps, use external tools, execute tasks and operate with reduced human oversight. They do not just respond to instructions. They pursue objectives.

For financial services, where trust, resilience and accountability are structural requirements, this does not represent an upgrade to the previous wave. It redefines what governance, risk exposure and professional capability require.

What This Means Across Four Dimensions

Exponential AI reshapes financial services simultaneously across four interconnected dimensions. Each one demands a response.

Work. The task and job layer is changing. Agentic AI can now execute the multi-step cognitive workflows that previously required skilled professionals: drafting credit assessments, synthesising regulatory filings, running scenario analyses.

The question is not whether your institution’s task structure will change. It is whether you have mapped which roles face the greatest exposure and what your people need to develop.

The Performance Hexagon, developed by Huy Nguyen Trieu, identifies the professionals most at risk as Task Robots: those who execute process without applying judgement. The transition to Problem Solver, System Thinker and ultimately Superstar is not optional.

Capability Infrastructure. Policy statements and pilot programmes are not enough.

Institutions need the governance architecture, oversight frameworks and workforce systems that allow them to deploy and learn from AI at scale. The CDE Innovation Prism provides the strategic lens for this work, asking not just what AI makes faster and cheaper, but what it makes genuinely new. Most institutions are still in efficiency mode. The third wave demands more.

Measurement. Most institutions know they need AI capability. Very few can define it with precision or measure it at scale.The CFTE AI Proficiency Framework defines three levels: AI Literacy (L1), Applied AI Practitioner (L2) and AI Systems and Decision Leader (L3),
across ten capability domains. Without a measurement standard, capability development is guesswork. The Exponential Preparedness Gap cannot be closed without first knowing how wide it is.

The Individual. The individual answer to the Exponential Preparedness Gap is the Supercharged Professional: someone who has deeply integrated AI capability into their professional practice, applies AI tools with judgement and continues developing as the capability landscape evolves. Not a one-time course. A professional orientation.

What Regulators Are Already Doing

The regulatory response to Mythos has been the fastest, most coordinated reaction to a technology development that financial supervisors have produced in years. The evidence is worth examining closely, because it signals where institutional expectations are heading.

The Financial Stability Board. Bank of England Governor Andrew Bailey, who chairs the FSB  the global risk watchdog coordinating financial rules for G20 economies  requested directly that Anthropic brief the FSB’s leading finance ministries and central banks on the cyber vulnerabilities Mythos had identified.

At a speech at Columbia University in April, Bailey warned the model could “crack the whole cyber risk world open.” The FSB confirmed it “welcomes engagement with Anthropic and other firms on emerging and frontier risks to global financial stability.” (Reuters, May 2026)

The ECB. On May 24, 2026, the European Central Bank convened an ad hoc emergency meeting rare by the ECB’s own standards specifically to address cybersecurity risks from the new generation of AI models. ECB Executive Board member Frank Elderson told the Financial Times: “This is something that is game-changing. We want banks to look into this seriously. The clock is ticking.” His framing was unambiguous: banks need to move from andante to presto. (FT,  May 2026)

The US Federal Reserve and Treasury. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an emergency meeting at Treasury’s Washington DC headquarters with the CEOs of the banks the Fed classifies as structurally important to the global financial system. The specific purpose was to ensure banks were prepared to defend against the cybersecurity risks identified by Mythos. (Bloomberg, April 2026)

Not being part of the response is not a neutral position. It is a risk that will now be visible to supervisors across multiple jurisdictions simultaneously.

The Question Every Leader in Financial Services Should Ask Today

Here is the honest question: if Mythos is the first signal, and three more capability jumps of comparable magnitude arrive in the next 18 months, is your institution prepared to govern them?

Not to stop them. To govern them. To understand what changes, to assess the exposure, to adapt the operating model and to maintain the trust of clients, regulators and the public that financial services depends on.

There are four things every leader should be able to answer right now:

One. Have we mapped which roles and tasks in our institution face the greatest exposure to the third wave? Or are we operating blind?

Two. Do we have a governance framework that is designed for AI systems that change faster than annual review cycles? Or are we applying model risk management designed for a static world to a dynamic one?

Three. Can we measure the AI proficiency of our workforce with precision? Do we know where the gaps are and what they cost us? Or are we guessing?

Four. Are our leaders and professionals oriented toward continuous capability development? Or do we treat AI learning as a one-time compliance activity?

If the honest answer to any of these is unclear, the gap is structural. And the time to close it is before the next signal arrives, not after.

Where to Start?

CFTE built this course because the industry needed it – Exponential AI and the Future of Financial Services, available without any paywall. No registration barrier.  

This was a deliberate decision. Exponential AI affects every level of financial services: bank leaders, regulators, risk and compliance teams, HR and L&D functions, technology teams, professionals at every stage of their career.

The more people inside financial institutions understand what is happening and why, the better the sector’s collective response will be. That outcome is too important to put behind a paywall.

The course is 90 minutes. It is taught by two of the most credible voices in this space.

Huy Nguyen Trieu, Co-Founder and CEO of CFTE, former Managing Director at Citi and Associate Fellow at Oxford Said, co-created one of the world’s largest Fintech courses, reaching 140,000 participants across 130 countries. He developed the frameworks referenced throughout this post: the CDE Innovation Prism, the Performance Hexagon and the AI Capability Engine. In the course, he builds the exponential capability evidence base, the Three Waves framework and the strategic lens for what institutions and professionals should do next.

Professor Douglas Arner, Kerry Holdings Professor in Law at HKU and Associate Director at Cambridge CCAF, is one of the foremost authorities on finance, technology and regulation. He addresses the regulatory and governance response to Exponential AI: what frameworks exist, where the gaps remain, and what practical steps firms and supervisors should take.

The course covers six areas:

• Why AI progress is exponential and what that means structurally;
• The Three Waves and where the third wave sits today;
• The CDE Innovation Prism applied to financial services strategy;
• How regulators are responding and where the institutional tensions remain;
• What leaders must do before the next capability jump arrives; and
• How individuals can begin the transition to becoming Supercharged Professionals.

It is designed to be used as a shared institutional briefing resource. If you are a leader, share it with your team. If you work in risk, compliance, HR or L&D, forward it to the people who need to understand this before your organisation’s next governance or strategy cycle.

The next Mythos moment is not a question of if. It is a question of when. The institutions and professionals who have done the foundational thinking will be the ones who respond with clarity rather than react with alarm.

• Start the free course here.
• Share the course with your organisation.

Further Reading

• CFTE AI Proficiency Framework — The measurement standard for AI proficiency in financial services
• Supercharged Professional Programme — The individual capability pathway for professionals navigating the third wave
• Senior Leadership Alignment Programme — Institutional alignment for leaders navigating Exponential AI
• AI Capability Diagnostic — Institutional assessment of AI readiness gaps across ten capability domains
• The AI-fication of Talents Whitepaper – CFTE’s foundational research on how AI transforms professional capability requirements

Structured Summary for Reference

What is Exponential AI in financial services? Exponential AI refers to the compounding acceleration of AI capability combined with the collapse in deployment cost, producing a pace of change that financial institutions, regulators and professionals were not structured to absorb. It is characterised by the arrival of agentic AI systems — such as Mythos — that can reason, plan and act autonomously.

What is the Exponential Preparedness Gap? The Exponential Preparedness Gap, a concept developed by CFTE Co-Founder Huy Nguyen Trieu, describes the widening distance between the pace of AI capability advancement and the speed at which institutions, regulators and professionals are adapting their governance, operating models, risk frameworks and skills.

What are the Three Waves of AI in financial services? The first wave was machine learning and prediction. The second wave was generative AI. The third wave, now arriving, is agentic and autonomous AI: systems that can reason across steps, use tools and act with limited oversight. Mythos is a signal of the third wave.

What is the CDE Innovation Prism? The CDE Innovation Prism is a strategic framework developed by Huy Nguyen Trieu that classifies AI impact as Cheaper/Better/Faster (C), Enhancing (E) or Different (D). It helps institutions move beyond efficiency thinking to assess what Exponential AI makes genuinely new.

CFTE works with more than 100 organisations including central banks, regulators, financial institutions and governments across 130+ countries. CFTE is AI Capability Infrastructure for financial services.

The post Mythos Got the Headlines. Exponential AI Is the Real Story Reshaping Financial Services. How to Prepare Before the Next Mythos Moment. appeared first on CFTE.

In early 2026, Mythos sent alarm bells across central banks, regulators, and financial institutions worldwide. Not because of how it became known, but because of what it could do.

Mythos was not designed as a cybersecurity tool. It was a general purpose model, the next step up from what millions of people use daily. Yet it could autonomously find and exploit vulnerabilities across virtually every major operating system and browser, without any human guidance, at a level exceeding the best human experts.

If a general purpose model could do this as a byproduct of its broader capabilities, what does the next generation look like?

Mythos is not the story. It is the signal.

---

Huy Nguyen Trieu

Co-founder, CFTE
Author, The AI-fication of Jobs
Former Managing Director, Citi

Prof. Douglas Arner

Kerry Holdings Professor in Law, University of Hong Kong
Associate Director, Cambridge Centre for Alternative Finance

Huy Nguyen Trieu and Professor Douglas Arner have spent decades working at the intersection of finance, technology, regulation, and innovation. Both believe the world is now entering a fundamentally new phase of AI. To help finance professionals understand what this shift means, and what to do about it, CFTE has made this course free.

---

The deeper issue

Exponential AI

Mythos points to something much larger. We have entered a third wave of AI, where systems can not only think and reason but act autonomously in the digital world. At the same time, AI is becoming ten to a hundred times more powerful and cheaper every year. What is impossible today could be possible tomorrow. Not tomorrow as a figure of speech. Tomorrow.

For financial services, this matters more than in almost any other sector. The industry is built on trust, risk management, regulation, and professional judgement. These are precisely the things exponential AI puts under pressure. Yet most firms and professionals are still responding incrementally. The sector had a decade to adapt to fintech. It has had roughly three years to move from AI to generative AI, to reasoning AI, to agentic AI. The pace will not slow down.

---

What the course covers

Huy opens by unpacking what Mythos revealed and why it signals something far larger. He introduces the three waves of AI, the concept of Exponential AI, and what it means for individuals and organisations to become anti-fragile in this environment.

Professor Arner then examines the implications for financial services specifically: why the sector was well positioned for the first wave, why that is no longer sufficient, and how different categories of risk, from malicious actors to infrastructure concentration to the loss of human expertise, require fundamentally different governance responses.

1.5 hours Self-paced Free

What you will learn

Key themes covered in the course

	What Mythos reveals about the third wave of AI and why it signals something much larger
	Why exponential AI makes linear planning obsolete and what a new strategic mindset looks like
	How to become a supercharged professional and why talent, not technology, is now the real differentiator
	The five categories of AI risk in financial services and why each requires a different governance response
	How institutions can move from AI-busy to AI-ready by rethinking their operating model

Start the course today

1.5 hours · Self-paced · Free

Enrol now →

The post Exponential AI and Financial Services: A Free Course appeared first on CFTE.